Consequences of constitutional choice: reflections on Tocqueville

3. Consequences of constitutional choice: reflections on Tocqueville JON ELSTER Tocqueville's Democracy in America is, among other things, an argument about the social consequences of constitutions. He draws attention to various features of the democratic constitution he observed in the United States, such as universal suffrage, the system of elected officials, the jury system and freedom of association and of expression. He then goes on to discuss how these institutions have various consequences for certain social values, such as prosperity, happiness and religious faith. His goal is clearly to evaluate democracy, as compared to other arrangements such as despotism, monarchy or aristocracy. For this he needs a method for tracing the full social effect of democratic institutions. It will not do to look at local effects, partial effects, short-term effects or transitional effects. Rather we must see the problem as one of general equilibrium. We must compare democracy as a going concern with other regimes also considered as going concerns. Tocqueville is not generally considered an important figure in the development of social science methodology. He wrote as an historian, with the historian's peculiar brand of arrogance, which is to make the theories and methods employed as unobtrusive as possible. They serve as scaffolding, useful in construction but not to be left visible in the finished work. Hence posterity has tended to focus on Tocqueville's substantive views, notably his theory of liberty and equality as the main values of modern societies, sometimes in harness with one another, sometimes in conflict. I shall attempt to show, however, that Tocqueville had a profound understanding - unequalled in his time, unsurpassed in ours - of the nature of social causation. I hope the demonstration will be of interest in itself, in addition to the 81 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 82 JON ELSTER implications it might have for constitutional design1 and for democratic theory.2 I Local versus global consequences Tocqueville warns us against two invalid inferences. First, from the fact that a proposition may be (locally) true with respect to any unit of analysis, we cannot conclude that it may be (globally) true of all units. To make this inference is to commit the fallacy of composition.3 Secondly, from the fact that a proposition is true when applied to all units, we cannot conclude that it remains valid when applied to one unit separately and exclusively. This inference would be an instance of the fallacy of division.4 An instructive example of the fallacy of composition is due to George Katona. For the individual businessman, the effect of a tax increase is the same as that of a rise in wages - it represents an extra cost that he will try to pass on to the consumer through higher prices. The local causal link is from higher taxes to higher prices. But this is valid only under ceteris paribus assumptions, since the effect of a general tax increase will be a reduction of aggregate demand and thus a fall in prices. The global causal link goes in the opposite direction.5 Tocqueville offers a similar analysis of the effect of marrying for love, a practice which is widespread in democracies. Here, as elsewhere, he was concerned with defending democracy against its rearguard critics: Our ancestors conceived a singular opinion with regard to marriage. As they had noticed that the few love matches which took place in their days almost always ended in tragedy, they came to the firm 1 Some of these implications I discuss in chapter 10 below. I make no pretension here to scholarship. I am engaged in a somewhat anachronistic dialogue with Tocqueville, in which I largely disregard the historical context in which he wrote and the various ambiguities that a close reading of the texts might discern. I do hope, of course, that Tocqueville scholars will find something here that speaks to their concern, but mine lie elsewhere. 3 See Elster (1978), pp. 97ff. fpr an analysis of this fallacy. 4 The fallacy of composition is the inference from "For any x, possibly x if F" to "Possibly for all x, x is F." The inverse inference is not fallacious, but valid. An invalid deduction, however, is the inference from "Possibly for all x, x is F" to "For any x, possibly x is the sole F." This I call the fallacy of division. I ought to add that this bears only a distant relation to the standard logical terminology (for which see Hamblin 1970). 5 Katona (1951), pp. 45ff. 2 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 83 conclusion that in such matters it was very dangerous to rely on one's own heart. They thought that chance saw clearer than choice. (596)6 Tocqueville goes on to point out two reasons why this view is untenable. First, to marry for love in a society where this is the exception is to court disaster, since going against the current tends to create hostility and in turn bitterness. Secondly, only very opinionated persons will go against the current in the first place - not a character feature conducive to happy marriages. The latter mechanism is a sampling effect, the former a real after-effect.7 They are both related to the necessarily exceptional act of going against the current something which it is possible for any one individual to do, but not for all. And Tocqueville concludes that There is therefore no just ground for surprise if, in an age of aristocracy, a man who chooses to consult nothing but his taste and inclination in selecting a wife soon finds that irregular morals and wretchedness break into his home life. But when such behaviour is part of the natural and usual order of things, when the social system makes it easy, when paternal authority supports it and public opinion recognizes it, one should not doubt that the internal peace of families will be increased thereby and conjugal faith better protected. (597) A more consequential instance of this fallacy (or a closely related one) is pointed out in the notes to the second volume of the Ancien regime, in the course of a discussion of the summons of the Estates-General in 1789. Why did the King, disastrously, agree to let the third estate have double representation and to have voting in common rather than by estates? True, this system already existed in the Languedoc, where it seemed to produce class harmony rather than class conflict. Yet, Tocqueville argues, the fatal mistake was that of not recognizing that "an institution which in one province only led to slight changes in the local constitution could not fail to disturb it violently and profoundly 6 7 Hereafter page references in the text are to Tocqueville (1969). For this distinction see Feller (1968), pp. 199ff. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 84 JON ELSTER the day it was applied to the whole nation."8 Note that this is not exactly the member-set or any-all fallacy identified above. Rather it is the part-whole fallacy of thinking that whatever is true of a smaller entity must also be true of a larger one. Tocqueville is not here pointing to the dangerous effects of extending the Languedoc system to all provinces, but to the risk of extending it to the national assembly, which is at a higher level than the provincial ones. Consider next the fallacy of division. It is an old axiom that "Qui peut le plus peut le moins." Whoever can lift ten pounds can also lift five. But there are cases in which the axiom is false, because in effect it embodies a fallacy of division. Thus an employer may be able to lay off all his workers, but unable to dismiss some but not all of his workers.9 Similarly, Tocqueville notes, in connection with the practice of compulsory military service, "a democratic government can do pretty well what it likes, provided that its orders apply to all and at the same moment; it is the inequality of a burden, not its weight, which usually provokes resistance" (651-2). This, in his opinion, is due not to solidarity but to envy, an endogenous vice of democracies. One may object, however, that the resistance could also be due to the lack of legitimacy of a decision that singles out a few to bear a burden that benefits all. There is, in all such cases, a bargaining problem that comes on the top of the collective action problem which the political system is supposed to resolve. People are more willing to forgo the possibility of being free riders than to see others benefiting without contributing - even when their additional contribution would be pointless. II Partial effects versus net effects When tracing the consequences of a given institutional change, it is all too easy to focus exclusively on one causal chain, forgetting that there may be several paths from the independent to the dependent variable. Moreover, even if there is only one primary causal chain, the effect 8 Tocqueville (1953), pp. 110-11. Nozick (1969), p. 480 refers to a ruling by the Supreme Court to the effect that "it is not an unfair labor practice for an employer to close his entire business, even if the closing is due to antiunion animus, but that closing part of his business is an unfair labor practice/' 9 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 85 may call forth countermeasures that partially or totally offset it.10 In both cases our interest ought to be in the net effect, rather than in any partial or ceteris paribus conclusions. An amusing example in Tocqueville is the following. "As there is no precautionary organization in the United States, there are more fires than in Europe, but generally they are put out more speedily, because the neighbors never fail to come quickly to the danger spot" (723). The general structure of this argument is the following. We want to examine the effect of the independent variable, democracy, on some dependent variable, such as the number of houses destroyed by fire. We first observe that the effect is mediated by two intermediating variables: the number of houses that catch fire and the proportion of fires that are not quickly extinguished. We also note that the two mediating variables interact multiplicatively in their effect on the dependent variable, rather than additively. Finally we observe that the first of the mediating variables is an increasing function of the independent variable, whereas the second is a decreasing one. This implies that the net effect could go either way, in the absence of more information about the relative strength of the two tendencies. Another example has to do with the impact of democracy on the strength of social interaction. Tocqueville argues that "the bonds of human affection are wider, but more relaxed" (507) than in aristocratic societies. Each person is tied to a greater number of other persons, but each tie is weaker. The substantive issue is whether what has been called "the strength of weak ties"11 exceeds or falls short of that of the closer relations in earlier societies. The methodological point, again, is that the question can be resolved only by looking at the net effect, rather than by focusing on some partial mechanism. A closely related form of argument is the following. Tocqueville first has the occasion to observe that democracy tends to increase people's opportunity set in some respect, and then goes on to point out 10 The distinction between these two cases is well brought out by considering Marx's theory of the tendency of the rate of profit to fall and the countertendencies that work in the opposite direction (Elster 1985, ch. 3). Some of the countertendencies are produced by the very same cause (i.e., labor-saving innovations) that generates the "main*' tendency, while others arise as reactions to the main tendency when businessmen see the rate of profit falling. 11 Granovetter (1973) argues that "weak ties play a role in effecting social cohesion" and that the "local cohesion of ethnic communities goes together with an extremely fragmented global society." Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 86 JON ELSTER that it simultaneously tends to weaken their desire to exploit the new opportunities. In the chapter in Democracy in America on "Why great revolutions will become rare," he argues that although men are less than satisfied with their present condition and feel no natural abhorrence toward revolution, they nevertheless are held back from it by their inclination. "The same social condition which prompts their longings restrains them within necessary limits. It gives them both greater freedom to change, and less interest in doing so" (636). Similarly he argues that the American Constitution "gave the president much power, but took away from him the will to use it" (138). The power stems from his prerogatives and veto, the lack of will to use it from the constant preoccupation with reelection. The same reasoning is applied to religion, as can be seen by juxtaposing two passages. "While the law allows the American people to do everything, there are things which religion prevents them from imagining and forbids them to dare" (292). And, "I doubt whether man can support complete religious independence and entire political liberty at the same time. I am led to think that if he has no faith he must obey, and if he is free he must believe" (444). Here the thrust of the second passage is that religion is endogenous to democracies, while the first argues that it tends to restrict the potentially dangerous freedom that is also part and parcel of democratic society. Usually Tocqueville uses this kind of argument to defend democracy against its critics, by suggesting that the net effect in question is positive, although the critic may correctly have perceived that one of the partial mechanisms by itself has bad consequences, other things being equal. Thus in democracies each thing is less well done, for reasons that will become clear, but this tendency is more than offset by the fact that more things are done (244). In particular, although democracies may perform badly in the short run, they out-perform aristocracies in the long term. Ill Short-term versus long-term consequences This distinction, indeed, is only a special case of the preceding one, but its importance in Tocqueville's analyses is such that it needs to be singled out for separate consideration. There is an interesting comparison that suggests itself here, between Tocqueville and SchumDownloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 87 peter. One obvious affinity is that they both attempt to work out a theory of capitalist democracy - Tocqueville for an early stage and Schumpeter for a later stage of that system. The relation between the two theories would deserve an extensive analysis, which would fall outside the scope of the present discussion. I want to draw attention, however, to the close structural resemblance between Schumpeter's defense of capitalism and Tocqueville's defense of democracy. Schumpeter admitted - in fact, insisted on - the multiple weakness of capitalism, such as its allocative inefficiency and its proneness to crises.12 He also argued that one should not judge capitalism as a system in the light of these defects, as they appear over a short span of time. The patent system or the excessively optimistic entrepreneurial expectations both involve short-term waste, but are also indispensable conditions for the dynamic efficiency of capitalism. In a famous passage Schumpeter makes the point in a very general manner: Since we are dealing with a process whose every element takes considerable time in revealing its true features and ultimate effects, there is no point in appraising the performance of that process ex visu of a given point of time; we must judge its performance over time, as it unfolds over decades or centuries. A system - any system, economic or other - that at every given point of time fully utilizes its possibilities to the best advantage may yet in the long run be inferior to a system that does so at no given point of time, because the latter's failure to do so may be a condition for the level or speed of long-run performance.13 In a very similar vein Tocqueville writes that "administrative centralization succeeds . . . in assembling, at a given time and place, all the available resources of the nation, but it militates against the increase of those resources" (88). Or again, "I think that in the long run government by democracy should increase the real forces of a society, but it cannot immediately assemble, at one point and at a given time, forces as great as those at the disposal of an aristocratic government or an absolute monarchy" (224). This holds in particular for warfare: "An aristocratic people which, fighting against a democ12 For discussion see Elster (1983), ch. 5. 13 Schumpeter (1961), p. 83. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 88 JON ELSTER racy, does not succeed in bringing it to ruin in the first campaign always runs a great risk of being defeated by it" (658). And if somebody objects to democracy on the grounds that it tends to tax the citizens too heavily, Tocqueville again has recourse to an argument of the same general form: Is democratic government economical? First, we must know with what we are comparing it. The question could easily be answered if we wanted to compare a democratic republic with an absolute monarchy. One would find public expenses in the former considerably greater than in the latter. But that is so of all free states compared with those not free. It is certain that despotism brings men to ruin more by preventing them from producing than by taking away the fount of wealth while often respecting acquired riches. But liberty engenders thousandfold more goods than it destroys, and in nations where it is understood, the people's resources always increase faster than the taxes. (208-9)14 IV Transitional effects versus steady-state effects The short-term versus long-term distinction must be kept separate from another distinction that needs to be made with respect to the temporal flow of consequences that stem from institutional change. This I refer to as the distinction between the transitional effects of introducing a certain institution, and the steady-state effect of having that institution. The latter can only be observed when all other institutions have adapted to the change, including all the chain effects and feedback effects that are set in motion. To bring out the difference between these two distinctions, we may first consider Tocqueville's discussion of freedom of association. As usual, he is concerned to defend this principle against the accusation that it perniciously undermines social stability. He admits that freedom of association might atfirstglance seem to involve a danger to society, but goes on to argue that through associations Americans also learn how to render the dangers of freedom less formidable: 14 Tocqueville does not, of course, argue that a high tax rate leads to a high taxable income. Rather, the argument is that democratic institutions are the common cause of both phenomena. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 89 By picking on one moment in the history of a nation it is easy to prove that political associations disturb the state and paralyze industry. But if you take the life of a people as one complete whole it may prove easy to show that freedom of political associations favors the welfare and even the tranquillity of the citizens. (534) This is the pure Schumpeterian argument of (III) above. But Tocqueville also makes a subtly different defense of associations when he argues "that political association is not nearly as dangerous to public peace as is supposed and that it could happen that it might give stability to a state after having shaken it for some time" (523). This, I believe, is a distinction between the temporary effect of introducing the freedom of association and the steady-state effect of having it. This reading might seem tenuous, but I believe it is plausible in view of the central importance of steady-state reasoning in the other passages that I now go on to discuss. Tocqueville is sympathetic to the usual conservative view that the stability of social arrangements is more important than their specific form. Speaking of the relation between master and servant, he observes that neither in stable aristocracies nor in stable democracies need there be anything degrading in either condition. Hence, he says, "It is not my business . . . to discover whether the new state of affairs which I have described is worse than what went before or simply different. It is enough for me that it is fixed and regulated, for what is important to find among men is not any particular order, but just order" (578). Correlatively, it is "in the journey from one social condition to the other" that "the lines between authority and tyranny, liberty and license" (579-80) become blurred. Either steady state, in other words, is preferable to the traverse from the one to the other. This argument, however, is not all there is to Tocqueville. Although notoriously ambivalent and ambiguous in his attitudes toward democracy, he does in many respects prefer - at least with his intellect, if not with his passions - the state of democracy over the old order. And so he is concerned to defend democracy by arguing that as a steady state it is superior to aristocracy and monarchy, although the process of democratization may initially lead from bad to worse. One step backward, two steps forward is the pattern of the democratic revolution. Hence, "One must be careful not to confuse the fact of equality Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 90 JON ELSTER with the revolution which succeeds in introducing it into the state of society and into the laws. In that lies the reason for almost all the phenomena which cause our surprise" (688). I shall briefly refer to some central passages in Democracy in America that apply this principle to central democratic institutions. (1) In his chapter on freedom of the press, Tocqueville argues that its steady-state effect is "to plunge mankind into universal doubt and distrust," while the transitional effect is to "daily change the object of their implicit belief." The former is a state of scepticism, the latter a series of firmly held and rapidly changing prejudices. That he clearly prefers the steady state to the traverse transpires from his warning: "Woe to these generations which first suddenly allow freedom to the press!" (187) Doubt is better than dogmatic belief, and although a justified conviction is better than either, it is rarely attained. (2) He also argues that "there is a tendency in democracy not to draw men together, but democratic revolutions make them run away from each other and perpetuate, in the midst of equality, hatreds originating in inequality. The Americans have this great advantage, that they attained democracy without the sufferings of a democratic revolution and that they were born equal instead of becoming so" (509). (3) As was pointed out above, he also asserts that freedom of political association is more dangerous during the traverse than in the steady state. (4) Moreover, "while equality favours sound morals, the social upheaval leading to it has a very damaging influence on them." In an argument that can be applied to the aftermath of the October Revolution, he also asserts that even those revolutions "which in the end imposed stricter moral standards began by relaxing them" (599). (5) Similarly, "although high ambitions swell while conditions are in process of equalization, that characteristic is lost when equality is a fact" (629). (6) Invoking the same argument, Tocqueville rejects the notion that "there must be a hidden and secret link between equality itself and revolutions" (634), since it is rather the process of equalization which involves violent and profound disturbances. (7) Lastly, and rather obviously, "the final result of a revolution might serve the interests of industry and trade, but its first effect will Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 91 almost always be the ruin of industrialists and traders, because it must immediately change general habits of consumption and temporarily upset the balance between supply and demand" (637). He might well have added the destructive effects on production. One could not wish, I think, for a more clearly drawn contrast. Whereas a society in the process of democratization and equalization fosters rapidly changing dogmatic beliefs, hostility between the citizens, ambition, loose morals, social unrest and economic ruin, a well-established democracy is in all these respects the opposite. 15 Although Tocqueville is no unconditional defender of democracy, he is acutely aware that one can evaluate it only as a going concern. Observe that the distinction between short-term suboptimality and long-term optimality is drawn within democracy as a going concern. Both are features of the steady state, so that the defects of the short term are not to be confused with those of the transitional period. Also, note the link between partial and transitional effects on the one hand, and between net effects and steady-effects on the other. The initial consequences of institutional change may be offset later on by consequences that take more time to work themselves out or are called forth in reaction to the immediate effects. Finally, note the difference between the presently discussed distinction and the distinction between local and global effects. To use Tocqueville's example, it could be the case that love matches continue to be unhappy even after they become the rule rather than the exception, until the process of adjustment has worked itself out fully. I add a few comments on Tocqueville's use of the distinction in other works. In the notes for the second volume of the Ancien regime he objects to the myopic character of Burke's Reflections on the Revolution in France. While Burke "is admirable when judging the details of the new institutions, their immediate consequences and the innumerable errors that resulted from the philosophical presumption and the lack of experience of the reformists," he is "only attentive to the forces that will be lost to France because of the Revolution, and not to those that will be gained." 16 Elsewhere he injects a note of 15 "The opposite" is, however, an ambiguous phrase. It may mean the absence of the phenomenon in question, or the presence of the substantively opposite phenomenon. The opposite of a system that makes people run away from one another may be one that does not have this effect or one that positively draws them together (Elster 1984a). 16 Tocqueville (1953), pp. 340-1. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 92 JON ELSTER scepticism concerning the character of those (steady-state) gains. "In my time I have already heard it proclaimed four times that the new society, such as the Revolution had made it, had finally found its natural and permanent state, and later on the next event proved that one was mistaken."17 This is actually a quite momentous observation. It suggests that there may not be any steady state at all toward which the chain of consequences converges, either because of the inherent character of the process or because of exogenous shocks that constantly change it in its course.18 There are hints of similar ideas in Democracy in America. Tocqueville believed that the American system he had observed contained potential for still further changes toward an aristocracy of money, a tyranny of the majority or a despotism founded on the equal subjection of all. Yet by and large he appeared to have believed that the state of society he observed around 1830 was "the natural and permanent" one, which might be contrasted with the immediate consequences of revolutionary upheaval. V Democracy and time A central substantive concern of Democracy in America is the ability of political systems to learn from the past and to engage in long-term planning. This issue involves time from the subjective point of view of the actors. The issues discussed under (III) and (IV) concern time from the point of view of the external observer. Perhaps the most striking achievement of Tocqueville's work is how he brings together attitudes toward time and performance over time. One might say, simplifying, that in his opinion Americans achieve so much in the long term because their political system is so badly adapted to long-term planning. A more correct statement would be that any reform designed to improve the capacity for long-term planning would have bad long-term consequences. Generally speaking, men are unique in that they can relate consciously to time - to events in the past and to possibilities in the future.19 They relate to the past mainly by learning and remembering. They relate to the future in a number of ways. The most basic is 17 Ibid., p. 343. I explore this distinction further in chapter 10. 19 This paragraph draws heavily on Elster (1984b). For useful surveys see also Crook (1980) and Staddon (1983). 18 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 93 grounded in the ability to wait, or to defer gratification. By an extension of this ability they can also deploy indirect strategies, of the form "one step backward, two steps forward." To say that men can behave in this way is not of course to say that they always do so when required. Impulsive behavior, or weakness of the will, is a constant threat to the ability to relate to the future. To neutralize the threat, it is sometimes possible to use a higher-order strategy, such as committing oneself in advance to a particular course of action from which one will be unable to deviate even should one later on want to do so. The problems created by one's inability to relate to the future may to some extent be alleviated by the ability to relate to that future inability. The air of paradox surrounding this statement is dissolved as soon as it is pointed out that precommitment today may require less will-power than resisting temptation will demand tomorrow. The preceding remarks apply to men individually, but one may also ask whether groups and organizations are good at relating to the past and the future. It is clear, for one thing, that groups with a very high turn-over rate will not be good at learning from past mistakes, or remembering past successes.20 Some continuity and overlap of membership is necessary. Similarly, if the tenure of managers and politicians is short, they will not be around to enjoy the benefits from innovations that do not yield immediate results. Hence with most types of incentive systems the organizations which they manage will have a short time horizon. I shall consider Tocqueville's views on these issues. Does democracy foster specific time attitudes in individuals? Does the democratic method for collective decision-making have particular consequences for the ability to learn from the past and take account of the future, over and above the effect on individuals? First, in democratic societies citizens tend to act on the assumption that techniques and preferences are in a constant state of flux, so that it is pointless to build anything too solid and durable. Tocqueville was apparently very impressed when he "met an American sailor and asked him why his country's ships are made so that they will not last long. He answered offhand that the art of navigation was making such quick progress that even the best of boats would be almost useless if it 20 Elster (1978), pp. 141ff. also develops this idea. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 94 JON ELSTER lasted more than a few years" (453). 21 Also, in a democracy people "are afraid of themselves, dreading that, their taste having changed, they will come to regret not being able to drop what once had formed the object of their lust" (582). 22 This in turn explains a fundamental fact about democracies: although people do more things than in aristocracies, they do each thing less well. "They carry through many undertakings quickly in preference to erecting long-lasting monuments" (631). This attitude to time contrasts greatly with that of the Ancients, who tended to look at the world as basically unchanging, so that it made sense to communicate with posterity by constructing durable monuments. 23 Either attitude may be rational, depending on the rate at which techniques and preferences do in fact change. Secondly, the incessant change of preferences is itself endogenous to democracies. "Americans cleave to the things of this world as if assured that they will never die, and yet are in such a rush to snatch any that come within their reach, as if expecting to stop living before they have relished them" (536). We find, therefore, "people continually changing path for fear of missing the shortest cut leading to happiness" (537). This echo of Descartes's famous argument24 is highly suggestive. Democracy does not foster the capacity for consistent long-term planning, because of an excessive concern with shortterm fine-tuning.25 Nor, thirdly, is democracy a good system for taking collective, future-oriented decisions. "A democracy finds it difficult to coordinate the details of a great undertaking and to fix on some plan and carry it through with determination in spite of obstacles. It has little capacity for combining measures in secret and waiting patiently for the 21 For a discussion of this idea see Rosenberg (1976). See also Cyert and de Groot (1975) and Goodin (1982), ch. 9 for this idea. For a discussion of the attitudes towards time in Classical Antiquity see Veyne (1976), pp. 642ff. 24 "Ma seconde maxime 6tait d'etre le plus ferme et le plus resolu en mes actions que je pourrais, et de ne suivre pas moins constamment les opinions les plus douteuses, lorsque je m'y serais un fois determine, que si elles eussent ete tres douteuses. Imitant en ceci les voyageurs qui, se trouvant egares en quelque foret, ne doivent pas errer en tournoyant tantot d'un cote, tantot d'un autre, ni encore moins s'arreter en une place, mais marcher toujours le plus droit qu'ils peuvent vers un meme cote", et ne le changer point pour de faibles raisons, encore que ce n'ait peut-etre 6te* au commencement que le hasard seul qui les a determine"s a le choisir; car, par ce moyen, s'ils ne vont justement ou ils d€sirent, ils arriveront au moins a la fin quelque part ou vraisemblablement ils seront mieux que dans le milieu d'une foret" (Descartes [1897-1910], vol. VI, p. 24). 25 Elster (1984b), ch. 1.4. 22 23 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 95 result" (229). Tocqueville, in other words, is making two different but strongly convergent arguments about the attitude toward time fostered by democracies. Individual as well as collective decisions suffer from an attitude toward the future that can variously be characterized as incontinent, inconstant or inconsistent.26 The fact of inconstancy, in particular, makes it rational not to invest too much in enterprises in which one may soon lose interest. Lastly, democratic governments have difficulties in relating to and learning from the past. Tocqueville explains this fact by the high turn-over rate which characterizes the American political system and social life generally. He saw clearly that the high rates of change and mobility in democratic societies have advantages as well as inconveniences. He argues, for instance, that since "the rich are constantly becoming poor or retiring from business when they have realized their profits," the "class of the rich does not exist at all." Hence the aristocracy of money "does not know its own mind and cannot act," and for that reason is less of a threat to democracy than an enduring group of privileged would be (557). But in the case of politics, the fact that "each generation is a new people" (473), unable to learn from the past, has uniformly bad consequences. "After one brief moment of power, officials are lost again amid the ever-changing crowd," and "nobody bothers about what was done before him." The upshot is that "It is very difficult for American administrators to learn anything from each other . . . So democracy, pressed to its ultimate limits, harms the progress of the art of government" (208). There seems to be an inconsistency in this part of Tocqueville's argument. He twice uses the phrase "the ability to make retrievable mistakes" to characterize the United States and to explain why democracy could succeed there and not elsewhere. In the first context he refers to the ability to learn by trial and error, to "profit by past experience" (225). But this is hard to reconcile with the argument about administrative instability which prevents learning from experience. Even in the first context, however, the argument is ambiguous, and he may have intended to say only what he definitely states in the 26 For incontinence, or ''weakness of the will," see Davidson (1980), ch. 2; for inconstancy, or endogenously changing preferences, see von Weiszacker (1971); for inconsistency, or the inability to stick to past plans because of non-exponential time preferences, see Ainslie (1975). Tocqueville's discussions are too casual to allow us to determine to which of these closely related phenomena he is referring. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 96 JON ELSTER second passage. Here he invokes the principle of the steady state, and argues that if a nation is so placed as to afford retrievable mistakes, it can survive the initial step backward of the transitional period. "Suppose a society so organized by nature or by constitution that it can tolerate the passing effect of bad laws and can without disaster await the result of the general tendency of its laws, and in such a case you will appreciate that democratic government, for all its faults, is yet the best suited of all to make society prosper" (232). I have referred twice to the notion of "one step backward, two steps forward" or, as Leibniz had it, "reculer pour mieux sauter."27 First, democracy as a going concern does not promote the ability to deploy such indirect strategies in a deliberate and conscious way, because of the peculiar attitude toward time it fosters. Secondly, in the transition to democracy the situation will tend to get worse before it gets better not as a result of strategic choice, but because of the general turmoil created by the lack of stable institutions. Tocqueville clearly saw that the fragility of the intermediate state might prevent thefinalstate from emerging, and suggested two conditions under which this difficulty might be overcome. In the case of America the country was less vulnerable because of its relative isolation and autarchy. The ability to make retrievable mistakes is linked to the fact that "no one needs the Americans, and they do not need anybody" (131). A second possibility is that of a simultaneous transition in several nations, so that no single nation can exploit the weaknesses of the others during the traverse. The "relative weakness of democratic republics in time of crisis is perhaps the greatest obstacle preventing the foundation of such a republic in Europe. For a democratic republic to survive without trouble in a European nation, it would be necessary for republics to be established in all the others at the same time" (224). How, then, shall we explain the fact that democracies prove their superiority in the long run even though they are badly suited for dealing with the future? I have already hinted at the general direction of Tocqueville's answer: democracy increases the forces of the nation, although it also causes them to be badly utilized. In the chapter on "The real advantages derived by American society from democratic government" Tocqueville elaborates as follows: 27 For further discussions of Leibniz see Elster (1984b), p. 10, n.19 and especially Elster (1975), pp. 21 Iff., pp. 229ff. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 97 That constantly renewed agitation introduced by democratic government into political life passes, then, into civil society. Perhaps, taking everything into consideration, that is the greatest advantage of democratic government, and I praise it much more on account of what it causes to be done than for what it does. It is incontestible that the people often manage public affairs very badly, but their concern therewith is bound to extend their mental horizon and shake them out of the rut of ordinary routine . . . Democracy does not provide a people with the most skillful of governments, but it does that which the most skillful government often cannot do: it spreads throughout the body social a restless activity, superabundant force and energy never found elsewhere, which, however little favoured by circumstances, can do wonders. Those are its true advantages. (243-4) The advantages of democracy, in other words, are mainly byproducts. The avowed aim of democracy is to be a good system of government, but Tocqueville argues that it does not realize this goal. Democratic governments lack the ability to proceed in a systematic, coherent way; they cannot plan for the future, or stick to past decisions. Yet the very activity of governing democratically has as its by-product a certain energy and restlessness that benefits industry and prosperity. A similar argument is advanced for the jury system: "I do not know whether a jury is useful to the litigants, but I am sure it is very good for those who have to decide the case. I regard it as one of the most effective means of popular education at society's disposal" (275). Again, the main justification for the institution is found in side-effects, which could hardly be the main motivation of the participants. (They could, of course, enter into the motivation of the creators of the system.) VI Conclusion Tocqueville's Democracy in America uses democracy as an independent variable to explain various features of American life: religion, public opinion, family life, economic activities, military matters and numerous other phenomena. The independent variable itself may be characterized as institutionalized equality and liberty. The equality is Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 98 JON ELSTER both social and political, i.e., it pertains both to the absence of privilege and hierarchy, and to the principle of universal suffrage. The liberty consists in the traditional political freedoms: freedom of religion, of association, of expression. Yet beyond these formal institutions democracy must also have a specific content. In particular, there must be not only the possibility of mobility, but a high degree of actual mobility between the social classes. The freedoms must be exercised, not just guaranteed. Hence the considerable amount of ambiguity in Tocqueville's references to democracy, and his own difficulties in distinguishing form from content.28 His intention, nevertheless, is fairly clear. Given that the democratic institutions exist and that the doors which they open are actually used, what are the further consequences for social life?29 This way of phrasing the issue poses a methodological problem: how can we distinguish cause from effect in stable democratic societies? Tocqueville's answer to this question had political as well as methodological benefits. In France the opponents of democracy were above all concerned with the abuses that followed in the wake of democratization. By arguing that these are not inherent in democracy as such, he offered a new brief to its defenders. At the same time he was enabled to justify his causal statements about the consequences of democracy. They appear as statements about the effect of a change in exogenous institutional variables upon the endogenous social variables. The change is measured by the difference between the old and the new equilibrium values of these variables, not by the difference that may be observed shortly after the change. Needless to say, this proposal begs several questions. For one thing, the distinction between the democratic form and the social content is not respected. The French abuses might have been due to the lack of the social concomitants of democracy, not to their merely transitional character. For another, the lack of any democratic revolution in America creates a difficulty for the argument that the French are only suffering from the revolution rather than from democracy as such. It could be - in 28 Lamberti (1983) is very useful in this regard. A n answer to this question would have the following form: If (p and q), then r. From this, of course, one cannot infer: If p , then (q and r). Yet one sometimes gets the impression that Tocqueville wants to use only the existence of the democratic institutions as the independent variable to explain both their observance and the more general phenomena of social life. 29 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 99 fact, Tocqueville himself makes this argument - that the virtues of American democracy are related to its relatively uncontested emergence, in which case the steady-state consequences of a revolutionary introduction could be quite different. In any case, however, the merits of the Tocquevillian methodology - as set out in (I) through (IV) above - are unaffected by these objections. Tocqueville - like his older contemporaries Chateaubriand and Stendhal - was a democrat by reason, an aristocrat by heart. In particular, he was very sensitive to the aristocratic virtues of individualism and perfectionism. Yet he had an answer to the aristocratic criticism of the mediocrity of all things democratic: democracy does more things (or causes more things to be done), although each thing is done less well. The democratic era is the age of quantity, after that of quality. This links up with a classical argument for democracy in the sense of majority rule rather than rule by the elite: "If quality is equal (or, as Hobbes more exactly put it, quality must be taken to be equal as a condition of peace) the only differentiating factor left is quantity." 30 Tocqueville adds an instrumental argument for democracy to this legitimacy-oriented argument. In the modern age, he seems to say, no one can stand up and proclaim for himself a superior political wisdom, however superior he might in fact be. Yet the admission of the many to the political system also galvanizes them into action, to the extent of offsetting the loss of excellence. Democracy is unavoidable: once mankind has eaten the fruits of equality, there is no way back. Yet, although inefficient as a system for decision-making, it can be justified by its non-political side-effects. Universal participation in politics is the price we have to pay if we want society to be energetic and prosperous; and in any case we do not have the choice. 30 Barry (1980), p. 193. Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 References Ainslie, G. 1975. Specious reward. Psychological Bulletin, 82, 463-96 Barry, B. 1980. Is democracy special? In P. Laslett and J.Fishkin (eds.), Philosophy, Politics and Society, Fifth Series, pp. 155-96. Oxford University Press Crook, J. 1980. The Evolution of Human Consciousness. Oxford University Press Cyert, R. M. and de Groot, M. H. 1975. Adaptive utility. In R. H. Day and T. Groves (eds.), Adaptive Economic Models, pp. 223-46. New York: Academic Press Davidson, D. 1980. Essays on Actions and Events. Oxford University Press Descartes, R. 1897-1910. Oeuvres completes, ed. C. Adam and P. Tannery. 11 vols. Paris: Vrin Elster, J. 1975. Leibniz et la formation de Vesprit capitaliste. Paris: Aubier-Montaigne 1978. Logic and Society. Chichester: Wiley 1983. Explaining Technical Change. Cambridge University Press 1984a. Active and passive negation: an essay in Ibanskian sociology. In P. Watzlawick (ed.), The Invented Reality, pp. 175-205. New York: Norton 1984b. Ulysses and the Sirens, rev. edn. Cambridge University Press 1985. Making Sense of Marx. Cambridge University Press Feller, W. 1968. An Introduction to Probability Theory, vol. 1, 3rd edn. New York: Wiley Goodin, R. 1982. Political Theory and Public Policy. Chicago University Press 1982 Granovetter, M. 1983. The strength of weak ties. American Journal of Sociology, 78, 1360-80 100 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 CONSEQUENCES OF CONSTITUTIONAL CHOICE 101 Hamblin, C. L. 1970. Fallacies. London: Methuen Katona, G. 1951. Psychological Analysis of Economic Behavior. New York: McGraw-Hill Lamberti, J.-C. 1983. Tocqueville et les deux democraties. Paris: Presses Universitaires de France Nozick, R. 1969. Coercion. In S. Morgenbasser et al. (eds.), Philosophy, Science and Method: Essays in Honor of Ernest Nagel, pp. 440-72. New York: St Martin's Press Rosenberg, N. 1976. On technological expectations. Economic Journal, 86, 525-35 Schumpeter, J. 1961. Capitalism, Socialism and Democracy. London: Allen and Unwin Staddon, J. E. R. 1983. Adaptive Behavior and Learning. Cambridge University Press Tocqueville, A. de. 1953. UAncien Regime et la revolution, vol. 2: Fragments et notes inedites sur la revolution. In Tocqueville, Oeuvres completes, vol. II. Paris: Gallimard 1969. Democracy in America. New York: Anchor Books Veyne, P. 1976. he Pain et le cirque. Paris: Le Seuil von Weiszacker, C. C. 1971. Notes on endogenous change of tastes. Journal of Economic Theory, 3, 345-72 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004 Downloaded from http:/www.cambridge.org/core. New York University Libraries, on 13 Dec 2016 at 17:28:20, subject to the Cambridge Core terms of use, available at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/CBO9781139173629.004